Tuesday, March 25, 2008

Ruthbert


My hilarious friend Ruth and I have been keeping each other company from our disparate, cross-city desk-jockey gigs for going on three years now. She works in search optimization.

Here, for posterity, are some of my recent conversations with her:


Ruth: I just accidentally sent an email asking someone for an "estimeat"

Ruth: :'(\
mr. stabby-face feels your pain
Ruth: he is sad because he has been stabbed in the face

K2: oh I hate getting lunch
I always feel guilty that I did not prepare my own 

K2: I currently do all my cooking at D's on the weekends
weeknights I live on prayer and the non-kosher noodle bowls I keep in my bedroom
Ruth: you should prepare your own lunch then!
[Ruth delivers office lunch recipe]

K2: I have noted your recipe and will take it up with my manager for possible inclusion in a future release
however this week's release is full, I'm sorry
Ruth: mission statement: to increase lunch-eater value by improving lunch efficiency and reducing lunch spend, while maintaining current levels of lunch deliciousness 

K2: of course you realize you're going to have to get signoff from the downstairs cafeteria lady, she's a stakeholder in this
and they're veeeery preoccupied with their chicken parmesean release right now, so good luck getting on THEIR radar
Ruth: I'm confident that the impact to downstairs cafeteria lady will be within acceptable bounds to her organization, and will be more than offset by the increased value offered by the new K2 lunch plan 

K2: Have you consulted EMEA? I'm sure that the Latin regions will buy in to the tortillas, but I don't know if you're aware of this, EMEA's servers are all allergic to gluten
Ruth: it wouldnt' be stored on EMEA's servers, it would be stored directly on K2's work fridge servers
and tortillas can be desk-hosted for several days before losing optimum freshness

Ruth: I am gradually accumulating Buffy seasons, when sale prices coincide with coupon-having on my part. M calls the shelf where I store them the "Garden of Whedon."
K2: 7 more bottles of meetings on the wall, 7 more bottles of meetings
cancel one down, send it around, 6 more bottles of meetings on the wall
Ruth: I hope that someone gets my
meeting in a bottle yeah


K2: I wrote a song on the way in to the tune of that "Shorty got low" song.
Here it is:
She had the powerpoint deck
And the laser poin-ter
The whole conference call was lookin' at her
Ruth: She hit her visio
next thing you know
spending got low, low, low, low low low
she had the Excel spreadsheet
and that ThinkPad in her lap
she turned around and gave that big budget a slap
Hey! She hit her visio
next thing you know
spending got low low low low low low

Ruth: in the last 6 weeks there have been 133 searches for the phrase "nose bidet"


Thursday, March 13, 2008

1.2 There is no "I" in "Meme": Language


The language we use, whether it beongs to our culture, subculture, technical tool or organization, affects our fundamental perceptions, choices, and behavior (and we rarely notice).

[T]he language a person speaks [affects] how that person both understands the world and behaves in it. [...] Different language patterns yield different patterns of thought. -- Sapir-Whorf Hypothesis, Wikipedia
The world, and the process of living, are inherently continuous and nameless. To communicate about them, we humans use our amazing minds to interpret them down into structures, finite bits with ends, beginnings and apparent affinities with other bits. We come up with those structures by drawing from our history, our cultural biases, our priorities, which themselves are created by language in an ongoing feedback loop. There is no "accurate" way of describing the world, and thus there is no objective way. There's just the way that you happen to be using.

A language is a very heavily biased cultural description of the parts of the world that it cares about, and its own priorities. For example:

In American culture (and some, but not all, other cultures), words about sex and sexual organs are used as the strongest forms of insults. It would be ridiculous for me to insult you by calling you a "stupid arm," but we find the word [cxxx] so offensive that I can't even directly reference it in this blog. This use of language reflects, and perpetuates, certain cultural assumptions about sex.

The languages of subcultures and organizations also codify, teach and perpetuate, unconscious cultural values and expectations.


A close friend of mine just became a cop. He is a deep and thoughtful person with a strong protective instinct. I was thus surprised, in recent conversations, to hear him talking about apprehending and charging people as "contacting" his "clients."
For me, "contacting a client" means calling someone who has voluntarily hired you and maybe leaving them, say, a voicemail, or perhaps a nice fruit basket. For his unit, "contacting" someone might mean chasing a guy four blocks, fighting him to the ground with the help of three other officers, finding a gun and a gram of coke on him, and taking him to jail where he will develop his first criminal record, permanently changing the course of his life.

Contrast this linguistic convention to some other options: instead of "Today I contacted four clients," make it "Today I handled, then made life-changing decisions towards, four human beings." Or, alternately: "Today, I wiped the floor with four more scumbags." The neutral language employed by my friend's unit serves two instructive purposes: it facilitates the emotional detachment that makes policework possible, and it restricts that detachment to the universe of professionalism and service.

So, yes: the linguistic structure you use effects how you think and act. But while changing a language is often a part of changing a social system's rules and behavior, it is rarely the direct route. If you had a goal, for example, to make all police highly sensitive to the human realities of the people they are arresting,* you might think it would be smart to change the institutional language from "contacting clients" to "Making Choices to affect Human Beings." But if your police still (for example) (1) have to fill an ambitious weekly quota of arrests to succeed at your unit, and (2) only see the perp at the time of the arrest (no exposure to context or after effects), your new language won't change their approach to the job. It will just irritate them as phony polish on top of the job's harsh reality.

I don't know about you, but I see surface-level language "fixes" in the industry all the time, and they drive me directly to drink.

To learn more about this fascinatng topic, you may want to read this highly-recommended book: Sorting Things Out: Classification and its Consequences.**

* And I hope you do not
** Lord knows I want to read it. It's on my short list.

Stay tuned for: 1.3 There is No "I" in "Meme": Culture

Tuesday, March 4, 2008

1.1 How to Build Horrible Social Systems by Accident: Incentives


In the last entry, I proposed that tools and organizations, often (make that usually-- actually make that almost always) unintentionally bake into their very structures a set of implicit instructions for their members / users about what behavior is appropriate, rewarded, or discouraged.

How, exactly, do they do that? Here are some of many possible answers.
Incentive Systems. Behaviors rewarded by your system or tool will grow in emphasis and frequency; behaviors that are punished will become less frequent. This statement may seem obvious; people are always trying to leverage positive or negative incentives to get one another to do things. Unfortunately, those conscious incentive programs are usually laid on top of preexisting incentive systems that are deeper, more subtle, more ubiquitous, and far less intentional. In other words, they are much more convincing to the people involved, and are impossible to casually override.

Example: You have a new software company, and you have to hire some people and then give them employee reviews of some kind. Like many orgs, you base your employee review system on whether or not an employee succeeds at his projects. If he succeeds at all of them, he gets a raise; if he fails at his projects, he gets a poor review and a lower bonus. If he gets three poor reviews in a row, he gets fired.
While this seems like a simple and obvious incentive system, you are literally incenting your average employee (let's call her Martha Generic) to succeed at her own projects… even if that messes up everyone else's. If she sacrifices her own project, one quarter, to enable four other projects to succeed, she will still be punished by your system.

Example, Cont'd: Five years down the line (after every manager in your company has worked your incentive system into dozens of mini-processes and deliverables), you discover your employees aren't collaborating. You say to yourself: "These poor geeks just don't know how to collaborate. I've got to get them thinking like a team…"

You start publishing some weekly articles on the importance of collaboration. You deliver a motivational speech to the whole company about how software development is really about putting "people first." You offer a trophy for the "most collaborative team member."

Will it work?

What would happen if you created an online community (let's say, a "resource group for workaholics") and let your members give each other public ratings (1-5 stars) on two things: "Humor" and "Best Vocabulary"? … What if it were an automated system that gave privileges based on "Most Links Contributed"?


Next up: 1.2 There is no I in Meme: Language and Messaging

1.1 How to Build Horrible Social Systems by Accident: Incentives


In the last entry, I proposed that tools and organizations, often (make that usually-- actually make that almost always) unintentionally bake into their very structures a set of implicit instructions for their members / users about what behavior is appropriate, rewarded, or discouraged.

How, exactly, do they do that? Here are some of many possible answers.
Incentive Systems. Behaviors rewarded by your system or tool will grow in emphasis and frequency; behaviors that are punished will become less frequent. This statement may seem obvious; people are always trying to leverage positive or negative incentives to get one another to do things. Unfortunately, those conscious incentive programs are usually laid on top of preexisting incentive systems that are deeper, more subtle, more ubiquitous, and far less intentional. In other words, they are much more convincing to the people involved, and are impossible to casually override.

Example: You have a new software company, and you have to hire some people and then give them employee reviews of some kind. Like many orgs, you base your employee review system on whether or not an employee succeeds at his projects. If he succeeds at all of them, he gets a raise; if he fails at his projects, he gets a poor review and a lower bonus. If he gets three poor reviews in a row, he gets fired.
While this seems like a simple and obvious incentive system, you are literally incenting your average employee (let's call her Martha Generic) to succeed at her own projects… even if that messes up everyone else's. If she sacrifices her own project, one quarter, to enable four other projects to succeed, she will still be punished by your system.

Example, Cont'd: Five years down the line (after every manager in your company has worked your incentive system into dozens of mini-processes and deliverables), you discover your employees aren't collaborating. You say to yourself: "These poor geeks just don't know how to collaborate. I've got to get them thinking like a team…"

You start publishing some weekly articles on the importance of collaboration. You deliver a motivational speech to the whole company about how software development is really about putting "people first." You offer a trophy for the "most collaborative team member."

Will it work?

What would happen if you created an online community (let's say, a "resource group for workaholics") and let your members give each other public ratings (1-5 stars) on two things: "Humor" and "Best Vocabulary"? … What if it were an automated system that gave privileges based on "Most Links Contributed"?


Next up: 1.2 There is no I in Meme: Language and Messaging